Multi-dimensional Images

A multi-dimensional image is a composite of several images of the same scene. The component images are registered so that equivalent pixels correspond to the same location on the ground.

A three-dimensional image is often displayed as a colour-composite in which the three individual images correspond to the red, green and blue components. Variations in colour reflect differences between the images caused by changes on the ground such as growth of vegetation, or change in water content.

Multi-dimensional images can be generated in various ways.

Complex Data.

The NetCDF image format treats complex data by storing the real and imaginary parts in separate layers; hence, formally a complex dataset is stored in a multi-dimensional image. We take this for granted below; but (for example) complex data is usually stored for the separate polarisations in a multi-polarimetric image.

Multi-temporal images.

The images are taken at different times, possibly months apart. This allows temporal changes to be monitored.

Multi-frequency images.

The images are formed with different frequency radars. Since details smaller than the wavelength of the beam are invisible, a multi-frequency image shows structures on different length scales.

Multi-polar images.

Like light, radar can be polarised into horizontal (H) and vertical (V) planes. Transmitting H or V polarised radiation and receiving H or V gives four possible images of the scene. Multi-polar images are also called polarimetric images.

Multi-sensor images.

Remote sensing images cover a large range of the electromagnetic spectrum: visible, infra-red, thermal radiation, and microwaves. Images from several bands can be selectively combined to discriminate between particular features of interest. InfoPACK makes it straightforward to carry out data-fusion processing in this way.